After years of building things quietly — training models, shipping agents, filing patents — I figured it’s time to start writing some of it down.
Why a blog?
I want to document what I am learning, the problems I face when adopting new technologies, and the solutions I figure out to overcome them. I also just want to share things I find interesting, obscure fixes I’ve stumbled upon, and ideas that might be helpful to others.
Currently, most of my thinking and research happens scattered across Obsidian notes, 100s of open browser tabs, and deep dives into obscure Reddit posts. A blog is a way to pull all of that chaos together to make that thinking durable and useful — for future me, and maybe for somebody else.
What I’ll write about
Mostly technical things I’m working on, learning, or thinking about:
- Computer vision — 3D image processing, detection, tracking pipelines, optimizing models for edge devices, and low-level image processing in C++.
- LLMs & Local Setup — setting up local models, trying them out, tinkering with inferencing, and understanding how to get the most out of the hardware I have.
- Engineering practice — solo engineering at a startup, prototyping fast, and knowing when to throw something away.
- Occasional opinion — on tooling, on research-to-product gaps, and on what “AI Engineer” actually means today.
I’m not aiming for polished essays. These will be more like annotated notes — rough around the edges, written to think, not to perform.
Setup
This site is built with Jekyll on GitHub Pages. Posts live as Markdown files in _posts/. The terminal aesthetic is hand-rolled CSS — no frameworks.
If something here is useful or wrong, you can reach me at simhadri1998@gmail.com.
# That's all for now
echo "More posts incoming..."